pacman::p_load(olsrr, corrplot, ggpubr, sf, spdep, GWmodel, tmap, tidyverse, glue, ggstatsplot, sfdep)In Class Exercise 7
7 Introduction
In this exercise, we will work on reinforcing our learning from Hands-on Exercise 7 with further exercises.
7.1 Importing the packages
The following packages are imported into our environment to facilitate analysis.
olsrr: Provides tools for building and evaluating Ordinary Least Squares (OLS) regression models, including diagnostic and selection methods.
corrplot: A package for visualizing correlation matrices using different methods, such as color-coded heatmaps and circles.
ggpubr: Facilitates easy creation of publication-ready plots based on ggplot2, with additional features for customization and statistical annotations.
sf: Stands for Simple Features, providing support for handling, analyzing, and visualizing spatial data within R.
spdep: Specializes in spatial dependence modeling and analysis, including spatial autocorrelation, spatial regression, and spatial weights generation.
GWmodel: A package that implements Geographically Weighted Regression (GWR) and other geographically weighted models for spatial data analysis.
tmap: Provides an intuitive syntax for creating thematic maps and handling spatial data, supporting both static and interactive maps.
tidyverse: A collection of R packages designed for data science that share a common philosophy, including data manipulation (dplyr), visualization (ggplot2), and more.
gtsummary: Simplifies the process of creating summary tables for statistical analyses, particularly useful for regression models and descriptive statistics.
The p_load() function of the pacman package is used as shown in the code chunk below.
7.2 Importing the data
7.2.1 Importing the geospatial data
We start off by importing the geospatial data into our environment. We use the st_read() function of the sf package for this.
mpsz = st_read(dsn = "data/geospatial", layer = "MP14_SUBZONE_WEB_PL")Reading layer `MP14_SUBZONE_WEB_PL' from data source
`C:\arjxn11\ISSS626-GAA\In-class_Ex\In-class_Ex7\data\geospatial'
using driver `ESRI Shapefile'
Simple feature collection with 323 features and 15 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 2667.538 ymin: 15748.72 xmax: 56396.44 ymax: 50256.33
Projected CRS: SVY21
This dataset is in ESRI shapefile format. The shapefile consists of URA Master Plan 2014’s planning subzone boundaries. Polygon features are used to represent these geographic boundaries. The GIS data is in svy21 projected coordinates systems. The geometry type is multipolygon.
We will now check the CRS information and update it if required.
EPSG code for Singapore is 3414.
We implement the st_crs() function of the sf package as shown in the code chunk below.
st_crs(mpsz)Coordinate Reference System:
User input: SVY21
wkt:
PROJCRS["SVY21",
BASEGEOGCRS["SVY21[WGS84]",
DATUM["World Geodetic System 1984",
ELLIPSOID["WGS 84",6378137,298.257223563,
LENGTHUNIT["metre",1]],
ID["EPSG",6326]],
PRIMEM["Greenwich",0,
ANGLEUNIT["Degree",0.0174532925199433]]],
CONVERSION["unnamed",
METHOD["Transverse Mercator",
ID["EPSG",9807]],
PARAMETER["Latitude of natural origin",1.36666666666667,
ANGLEUNIT["Degree",0.0174532925199433],
ID["EPSG",8801]],
PARAMETER["Longitude of natural origin",103.833333333333,
ANGLEUNIT["Degree",0.0174532925199433],
ID["EPSG",8802]],
PARAMETER["Scale factor at natural origin",1,
SCALEUNIT["unity",1],
ID["EPSG",8805]],
PARAMETER["False easting",28001.642,
LENGTHUNIT["metre",1],
ID["EPSG",8806]],
PARAMETER["False northing",38744.572,
LENGTHUNIT["metre",1],
ID["EPSG",8807]]],
CS[Cartesian,2],
AXIS["(E)",east,
ORDER[1],
LENGTHUNIT["metre",1,
ID["EPSG",9001]]],
AXIS["(N)",north,
ORDER[2],
LENGTHUNIT["metre",1,
ID["EPSG",9001]]]]
We note that the current EPSG code is 9001, which is inaccurate. We must update this to 3414. The st_transform() function of the sf package will be implemented.
mpsz_svy21=st_transform(mpsz, 3414)
st_crs(mpsz_svy21)Coordinate Reference System:
User input: EPSG:3414
wkt:
PROJCRS["SVY21 / Singapore TM",
BASEGEOGCRS["SVY21",
DATUM["SVY21",
ELLIPSOID["WGS 84",6378137,298.257223563,
LENGTHUNIT["metre",1]]],
PRIMEM["Greenwich",0,
ANGLEUNIT["degree",0.0174532925199433]],
ID["EPSG",4757]],
CONVERSION["Singapore Transverse Mercator",
METHOD["Transverse Mercator",
ID["EPSG",9807]],
PARAMETER["Latitude of natural origin",1.36666666666667,
ANGLEUNIT["degree",0.0174532925199433],
ID["EPSG",8801]],
PARAMETER["Longitude of natural origin",103.833333333333,
ANGLEUNIT["degree",0.0174532925199433],
ID["EPSG",8802]],
PARAMETER["Scale factor at natural origin",1,
SCALEUNIT["unity",1],
ID["EPSG",8805]],
PARAMETER["False easting",28001.642,
LENGTHUNIT["metre",1],
ID["EPSG",8806]],
PARAMETER["False northing",38744.572,
LENGTHUNIT["metre",1],
ID["EPSG",8807]]],
CS[Cartesian,2],
AXIS["northing (N)",north,
ORDER[1],
LENGTHUNIT["metre",1]],
AXIS["easting (E)",east,
ORDER[2],
LENGTHUNIT["metre",1]],
USAGE[
SCOPE["Cadastre, engineering survey, topographic mapping."],
AREA["Singapore - onshore and offshore."],
BBOX[1.13,103.59,1.47,104.07]],
ID["EPSG",3414]]
7.2.2 Importing and wrangling the Aspatial Data
Since this is in CSV format, we implement read_csv() of the readr package to import it.
Always be careful to use read_csv() rather than read.csv().
condo_resale = read_csv("data/aspatial/Condo_resale_2015.csv")After importing the data file into R, it is important for us to examine if the data file has been imported correctly.
The codes chunks below uses glimpse() to display the data structure.
glimpse(condo_resale)Rows: 1,436
Columns: 23
$ LATITUDE <dbl> 1.287145, 1.328698, 1.313727, 1.308563, 1.321437,…
$ LONGITUDE <dbl> 103.7802, 103.8123, 103.7971, 103.8247, 103.9505,…
$ POSTCODE <dbl> 118635, 288420, 267833, 258380, 467169, 466472, 3…
$ SELLING_PRICE <dbl> 3000000, 3880000, 3325000, 4250000, 1400000, 1320…
$ AREA_SQM <dbl> 309, 290, 248, 127, 145, 139, 218, 141, 165, 168,…
$ AGE <dbl> 30, 32, 33, 7, 28, 22, 24, 24, 27, 31, 17, 22, 6,…
$ PROX_CBD <dbl> 7.941259, 6.609797, 6.898000, 4.038861, 11.783402…
$ PROX_CHILDCARE <dbl> 0.16597932, 0.28027246, 0.42922669, 0.39473543, 0…
$ PROX_ELDERLYCARE <dbl> 2.5198118, 1.9333338, 0.5021395, 1.9910316, 1.121…
$ PROX_URA_GROWTH_AREA <dbl> 6.618741, 7.505109, 6.463887, 4.906512, 6.410632,…
$ PROX_HAWKER_MARKET <dbl> 1.76542207, 0.54507614, 0.37789301, 1.68259969, 0…
$ PROX_KINDERGARTEN <dbl> 0.05835552, 0.61592412, 0.14120309, 0.38200076, 0…
$ PROX_MRT <dbl> 0.5607188, 0.6584461, 0.3053433, 0.6910183, 0.528…
$ PROX_PARK <dbl> 1.1710446, 0.1992269, 0.2779886, 0.9832843, 0.116…
$ PROX_PRIMARY_SCH <dbl> 1.6340256, 0.9747834, 1.4715016, 1.4546324, 0.709…
$ PROX_TOP_PRIMARY_SCH <dbl> 3.3273195, 0.9747834, 1.4715016, 2.3006394, 0.709…
$ PROX_SHOPPING_MALL <dbl> 2.2102717, 2.9374279, 1.2256850, 0.3525671, 1.307…
$ PROX_SUPERMARKET <dbl> 0.9103958, 0.5900617, 0.4135583, 0.4162219, 0.581…
$ PROX_BUS_STOP <dbl> 0.10336166, 0.28673408, 0.28504777, 0.29872340, 0…
$ NO_Of_UNITS <dbl> 18, 20, 27, 30, 30, 31, 32, 32, 32, 32, 34, 34, 3…
$ FAMILY_FRIENDLY <dbl> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0…
$ FREEHOLD <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1…
$ LEASEHOLD_99YR <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
head(condo_resale$LONGITUDE) #see the data in XCOORD column[1] 103.7802 103.8123 103.7971 103.8247 103.9505 103.9386
head(condo_resale$LATITUDE) #see the data in YCOORD column[1] 1.287145 1.328698 1.313727 1.308563 1.321437 1.314198
We now implement the summary() function of base R to condo_resale.
summary(condo_resale) LATITUDE LONGITUDE POSTCODE SELLING_PRICE
Min. :1.240 Min. :103.7 Min. : 18965 Min. : 540000
1st Qu.:1.309 1st Qu.:103.8 1st Qu.:259849 1st Qu.: 1100000
Median :1.328 Median :103.8 Median :469298 Median : 1383222
Mean :1.334 Mean :103.8 Mean :440439 Mean : 1751211
3rd Qu.:1.357 3rd Qu.:103.9 3rd Qu.:589486 3rd Qu.: 1950000
Max. :1.454 Max. :104.0 Max. :828833 Max. :18000000
AREA_SQM AGE PROX_CBD PROX_CHILDCARE
Min. : 34.0 Min. : 0.00 Min. : 0.3869 Min. :0.004927
1st Qu.:103.0 1st Qu.: 5.00 1st Qu.: 5.5574 1st Qu.:0.174481
Median :121.0 Median :11.00 Median : 9.3567 Median :0.258135
Mean :136.5 Mean :12.14 Mean : 9.3254 Mean :0.326313
3rd Qu.:156.0 3rd Qu.:18.00 3rd Qu.:12.6661 3rd Qu.:0.368293
Max. :619.0 Max. :37.00 Max. :19.1804 Max. :3.465726
PROX_ELDERLYCARE PROX_URA_GROWTH_AREA PROX_HAWKER_MARKET PROX_KINDERGARTEN
Min. :0.05451 Min. :0.2145 Min. :0.05182 Min. :0.004927
1st Qu.:0.61254 1st Qu.:3.1643 1st Qu.:0.55245 1st Qu.:0.276345
Median :0.94179 Median :4.6186 Median :0.90842 Median :0.413385
Mean :1.05351 Mean :4.5981 Mean :1.27987 Mean :0.458903
3rd Qu.:1.35122 3rd Qu.:5.7550 3rd Qu.:1.68578 3rd Qu.:0.578474
Max. :3.94916 Max. :9.1554 Max. :5.37435 Max. :2.229045
PROX_MRT PROX_PARK PROX_PRIMARY_SCH PROX_TOP_PRIMARY_SCH
Min. :0.05278 Min. :0.02906 Min. :0.07711 Min. :0.07711
1st Qu.:0.34646 1st Qu.:0.26211 1st Qu.:0.44024 1st Qu.:1.34451
Median :0.57430 Median :0.39926 Median :0.63505 Median :1.88213
Mean :0.67316 Mean :0.49802 Mean :0.75471 Mean :2.27347
3rd Qu.:0.84844 3rd Qu.:0.65592 3rd Qu.:0.95104 3rd Qu.:2.90954
Max. :3.48037 Max. :2.16105 Max. :3.92899 Max. :6.74819
PROX_SHOPPING_MALL PROX_SUPERMARKET PROX_BUS_STOP NO_Of_UNITS
Min. :0.0000 Min. :0.0000 Min. :0.001595 Min. : 18.0
1st Qu.:0.5258 1st Qu.:0.3695 1st Qu.:0.098356 1st Qu.: 188.8
Median :0.9357 Median :0.5687 Median :0.151710 Median : 360.0
Mean :1.0455 Mean :0.6141 Mean :0.193974 Mean : 409.2
3rd Qu.:1.3994 3rd Qu.:0.7862 3rd Qu.:0.220466 3rd Qu.: 590.0
Max. :3.4774 Max. :2.2441 Max. :2.476639 Max. :1703.0
FAMILY_FRIENDLY FREEHOLD LEASEHOLD_99YR
Min. :0.0000 Min. :0.0000 Min. :0.0000
1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
Median :0.0000 Median :0.0000 Median :0.0000
Mean :0.4868 Mean :0.4227 Mean :0.4882
3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:1.0000
Max. :1.0000 Max. :1.0000 Max. :1.0000
7.2.2.1 Converting aspatial data frame into a sf object
Currently, the condo_resale data frame is aspatial. We will convert it to a sf object. The code chunk below converts condo_resale data frame into a simple feature data frame by using st_as_sf() function of sf package.
condo_resale.sf <- st_as_sf(condo_resale,
coords = c("LONGITUDE", "LATITUDE"),
crs=4326) %>%
st_transform(crs=3414)Notice that st_transform() of sf package is used to convert the coordinates from wgs84 (i.e. crs:4326) to svy21 (i.e. crs=3414). We first set the CRS to 4326 as CSV data does not have the projection information.
Next, head() is used to list the content of condo_resale.sf object.
head(condo_resale.sf)Simple feature collection with 6 features and 21 fields
Geometry type: POINT
Dimension: XY
Bounding box: xmin: 22085.12 ymin: 29951.54 xmax: 41042.56 ymax: 34546.2
Projected CRS: SVY21 / Singapore TM
# A tibble: 6 × 22
POSTCODE SELLING_PRICE AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 118635 3000000 309 30 7.94 0.166 2.52
2 288420 3880000 290 32 6.61 0.280 1.93
3 267833 3325000 248 33 6.90 0.429 0.502
4 258380 4250000 127 7 4.04 0.395 1.99
5 467169 1400000 145 28 11.8 0.119 1.12
6 466472 1320000 139 22 10.3 0.125 0.789
# ℹ 15 more variables: PROX_URA_GROWTH_AREA <dbl>, PROX_HAWKER_MARKET <dbl>,
# PROX_KINDERGARTEN <dbl>, PROX_MRT <dbl>, PROX_PARK <dbl>,
# PROX_PRIMARY_SCH <dbl>, PROX_TOP_PRIMARY_SCH <dbl>,
# PROX_SHOPPING_MALL <dbl>, PROX_SUPERMARKET <dbl>, PROX_BUS_STOP <dbl>,
# NO_Of_UNITS <dbl>, FAMILY_FRIENDLY <dbl>, FREEHOLD <dbl>,
# LEASEHOLD_99YR <dbl>, geometry <POINT [m]>
Notice that the output is in point feature data frame. This is because it has only latitude and longitude.
7.3 Exploratory Data Analysis
7.3.1 EDA using statistical graphics
We can plot the distribution of SELLING_PRICE by using appropriate Exploratory Data Analysis (EDA) as shown in the code chunk below.
ggplot(data=condo_resale.sf, aes(x=`SELLING_PRICE`)) + geom_histogram(bins=20, color="black", fill="light blue")
The figure above reveals a right skewed distribution. This means that more condominium units were transacted at relative lower prices.
Statistically, the skewed dsitribution can be normalised by using log transformation. The code chunk below is used to derive a new variable called LOG_SELLING_PRICE by using a log transformation on the variable SELLING_PRICE. It is performed using mutate() of dplyr package.
condo_resale.sf <- condo_resale.sf %>% mutate(`LOG_SELLING_PRICE` = log(SELLING_PRICE))Now, you can plot the LOG_SELLING_PRICE using the code chunk below.
ggplot(data=condo_resale.sf, aes(x=`LOG_SELLING_PRICE`)) + geom_histogram(bins=20, color="black", fill="light blue")
Notice that the distribution is relatively less skewed after the transformation.
7.3.2 Multiple Histogram Plots distribution of variables
We now plot multiple histograms (also known as trellis plot) by using the ggarrange() function of the ggpubr package.
The code chunk below is used to create 12 histograms. Then, ggarrange() is used to organised these histogram into a 3 columns by 4 rows small multiple plot.
AREA_SQM <- ggplot(data=condo_resale.sf, aes(x= `AREA_SQM`)) + geom_histogram(bins=20, color="black", fill="light blue")
AGE <- ggplot(data=condo_resale.sf, aes(x= `AGE`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_CBD <- ggplot(data=condo_resale.sf, aes(x= `PROX_CBD`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_CHILDCARE <- ggplot(data=condo_resale.sf, aes(x= `PROX_CHILDCARE`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_ELDERLYCARE <- ggplot(data=condo_resale.sf, aes(x= `PROX_ELDERLYCARE`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_URA_GROWTH_AREA <- ggplot(data=condo_resale.sf, aes(x= `PROX_URA_GROWTH_AREA`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_HAWKER_MARKET <- ggplot(data=condo_resale.sf, aes(x= `PROX_HAWKER_MARKET`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_KINDERGARTEN <- ggplot(data=condo_resale.sf, aes(x= `PROX_KINDERGARTEN`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_MRT <- ggplot(data=condo_resale.sf, aes(x= `PROX_MRT`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_PARK <- ggplot(data=condo_resale.sf, aes(x= `PROX_PARK`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_PRIMARY_SCH <- ggplot(data=condo_resale.sf, aes(x= `PROX_PRIMARY_SCH`)) + geom_histogram(bins=20, color="black", fill="light blue")
PROX_TOP_PRIMARY_SCH <- ggplot(data=condo_resale.sf, aes(x= `PROX_TOP_PRIMARY_SCH`)) + geom_histogram(bins=20, color="black", fill="light blue")
ggarrange(AREA_SQM, AGE, PROX_CBD, PROX_CHILDCARE, PROX_ELDERLYCARE, PROX_URA_GROWTH_AREA, PROX_HAWKER_MARKET, PROX_KINDERGARTEN, PROX_MRT, PROX_PARK, PROX_PRIMARY_SCH, PROX_TOP_PRIMARY_SCH, ncol = 3, nrow = 4)
7.3.3 Drawing a statistical point map
Lastly, we want to reveal the geospatial distribution condominium resale prices in Singapore. The map will be prepared by using tmap package.
tmap_mode("view")
tm_shape(mpsz_svy21)+
tm_polygons() +
tm_shape(condo_resale.sf) +
tm_dots(col = "SELLING_PRICE", alpha = 0.6,style="quantile")+
tm_view(set.zoom.limits = c(11,14))+
tmap_options(check.and.fix = TRUE)We change tmap_mode back to plot before proceeding.
tmap_mode('plot')7.4 Hedonic Pricing Modelling
Hedonic pricing modeling is an econometric technique used to estimate the value of a good or service by breaking down the price into its component attributes. Commonly applied in real estate, it involves analyzing how individual factors such as location, size, amenities, or proximity to schools influence the overall market price of a property. This model helps in understanding how much each characteristic contributes to the price, separating the effect of specific features from the overall value.
We implement the lm() function of base R to build hedonic pricing models for condominium resale units.
7.4.1 Simple Linear Regression Method
First, we will build a simple linear regression model by using SELLING_PRICE as the dependent variable and AREA_SQM as the independent variable.
condo.slr <- lm(formula=SELLING_PRICE ~ AREA_SQM, data = condo_resale.sf)lm() returns an object of class “lm” or for multiple responses of class c(“mlm”, “lm”).
The functions summary() and anova() can be used to obtain and print a summary and analysis of variance table of the results. The generic accessor functions coefficients, effects, fitted.values and residuals extract various useful features of the value returned by lm.
summary(condo.slr)
Call:
lm(formula = SELLING_PRICE ~ AREA_SQM, data = condo_resale.sf)
Residuals:
Min 1Q Median 3Q Max
-3695815 -391764 -87517 258900 13503875
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -258121.1 63517.2 -4.064 5.09e-05 ***
AREA_SQM 14719.0 428.1 34.381 < 2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 942700 on 1434 degrees of freedom
Multiple R-squared: 0.4518, Adjusted R-squared: 0.4515
F-statistic: 1182 on 1 and 1434 DF, p-value: < 2.2e-16
The R-squared value of 0.4518 indicates that the simple regression model explains approximately 45% of the variation in resale prices.
Given that the p-value is much smaller than 0.0001, we can confidently reject the null hypothesis that the mean is an adequate predictor of the SELLING_PRICE. This suggests that the simple linear regression model is a significantly better estimator of SELLING_PRICE.
The Coefficients section of the report shows that the p-values for both the Intercept and ARA_SQM estimates are less than 0.001. This allows us to reject the null hypothesis that B0 (the intercept) and B1 (the slope for ARA_SQM) are equal to zero. Therefore, we can conclude that B0 and B1 are reliable parameter estimates.
To visualize the best fit line on a scatterplot, we can use the lm() method within ggplot’s geometry functions, as demonstrated in the following code snippet.
ggplot(data=condo_resale.sf, aes(x=`AREA_SQM`, y=`SELLING_PRICE`)) + geom_point() + geom_smooth(method = lm)
The figure above reveals that there are indeed a few statistical outliers with relatively high selling prices.
7.4.2 Multiple Linear Regression Method
7.4.2.1 Visualising the relationships of the independent variables
Before building a multiple regression model, it is important to ensure that the indepdent variables used are not highly correlated to each other. If these highly correlated independent variables are used in building a regression model by mistake, the quality of the model will be compromised. This phenomenon is known as multicollinearity in statistics.
Correlation matrix is commonly used to visualise the relationships between the independent variables. Beside the pairs() of R, there are many packages support the display of a correlation matrix. In this section, the corrplot package will be used.
The code chunk below is used to plot a scatterplot matrix of the relationship between the independent variables in condo_resale data.frame.
corrplot(cor(condo_resale[, 5:23]),
diag = FALSE, order = "AOE",
tl.pos = "td", tl.cex = 0.5,
method = "number", type = "upper")
Matrix reorder is very important for mining the hiden structure and patter in the matrix. There are four methods in corrplot (parameter order), named “AOE”, “FPC”, “hclust”, “alphabet”. In the code chunk above, AOE order is used. It orders the variables by using the angular order of the eigenvectors method suggested by Michael Friendly.
From the scatterplot matrix, it is clear that Freehold is highly correlated to LEASE_99YEAR. In view of this, it is wiser to only include either one of them in the subsequent model building. As a result, LEASE_99YEAR is excluded in the subsequent model building.
ggcorrmat(condo_resale[, 5:23])
7.4.3 Building a hedonic pricing model using multiple linear regression method
The code chunk below using lm() to calibrate the multiple linear regression model.
Note that we have added an additional variable LEASEHOLD_99YR to the below model for this in-class exercise.
condo.mlr <- lm(formula = SELLING_PRICE ~ AREA_SQM + AGE+
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_HAWKER_MARKET + PROX_KINDERGARTEN +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_TOP_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_SUPERMARKET +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD+ LEASEHOLD_99YR,
data=condo_resale.sf)
summary(condo.mlr)
Call:
lm(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD + PROX_CHILDCARE +
PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA + PROX_HAWKER_MARKET +
PROX_KINDERGARTEN + PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_TOP_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_SUPERMARKET +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD +
LEASEHOLD_99YR, data = condo_resale.sf)
Residuals:
Min 1Q Median 3Q Max
-3471036 -286903 -22426 239412 12254549
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 543071.4 136210.9 3.987 7.03e-05 ***
AREA_SQM 12688.7 370.1 34.283 < 2e-16 ***
AGE -24566.0 2766.0 -8.881 < 2e-16 ***
PROX_CBD -78122.0 6791.4 -11.503 < 2e-16 ***
PROX_CHILDCARE -333219.0 111020.3 -3.001 0.002734 **
PROX_ELDERLYCARE 170950.0 42110.8 4.060 5.19e-05 ***
PROX_URA_GROWTH_AREA 38507.6 12523.7 3.075 0.002147 **
PROX_HAWKER_MARKET 23801.2 29299.9 0.812 0.416739
PROX_KINDERGARTEN 144098.0 82738.7 1.742 0.081795 .
PROX_MRT -322775.9 58528.1 -5.515 4.14e-08 ***
PROX_PARK 564487.9 66563.0 8.481 < 2e-16 ***
PROX_PRIMARY_SCH 186170.5 65515.2 2.842 0.004553 **
PROX_TOP_PRIMARY_SCH -477.1 20598.0 -0.023 0.981525
PROX_SHOPPING_MALL -207721.5 42855.5 -4.847 1.39e-06 ***
PROX_SUPERMARKET -48074.7 77145.3 -0.623 0.533273
PROX_BUS_STOP 675755.0 138552.0 4.877 1.20e-06 ***
NO_Of_UNITS -216.2 90.3 -2.394 0.016797 *
FAMILY_FRIENDLY 142128.3 47055.1 3.020 0.002569 **
FREEHOLD 300646.5 77296.5 3.890 0.000105 ***
LEASEHOLD_99YR -77137.4 77570.9 -0.994 0.320192
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 755800 on 1416 degrees of freedom
Multiple R-squared: 0.652, Adjusted R-squared: 0.6474
F-statistic: 139.6 on 19 and 1416 DF, p-value: < 2.2e-16
7.4.4 Preparing Publication Quality Table: olsrr method
With reference to the report above, it is clear that not all the independent variables are statistically significant. We will revised the model by removing those variables which are not statistically significant.
Now, we are ready to calibrate the revised model by using the code chunk below.
condo.mlr1 <- lm(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_MRT + PROX_PARK +
PROX_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD+ LEASEHOLD_99YR,
data=condo_resale.sf)
ols_regress(condo.mlr1) Model Summary
-----------------------------------------------------------------------------
R 0.807 RMSE 751634.509
R-Squared 0.651 MSE 571320119490.610
Adj. R-Squared 0.647 Coef. Var 43.162
Pred R-Squared 0.637 AIC 42967.367
MAE 413020.461 SBC 43056.951
-----------------------------------------------------------------------------
RMSE: Root Mean Square Error
MSE: Mean Square Error
MAE: Mean Absolute Error
AIC: Akaike Information Criteria
SBC: Schwarz Bayesian Criteria
ANOVA
--------------------------------------------------------------------------------
Sum of
Squares DF Mean Square F Sig.
--------------------------------------------------------------------------------
Regression 1.513372e+15 15 1.008915e+14 176.594 0.0000
Residual 8.112746e+14 1420 571320119490.610
Total 2.324647e+15 1435
--------------------------------------------------------------------------------
Parameter Estimates
-----------------------------------------------------------------------------------------------------------------
model Beta Std. Error Std. Beta t Sig lower upper
-----------------------------------------------------------------------------------------------------------------
(Intercept) 591539.643 121110.937 4.884 0.000 353964.068 829115.218
AREA_SQM 12754.325 367.962 0.582 34.662 0.000 12032.517 13476.133
AGE -24822.087 2756.860 -0.168 -9.004 0.000 -30230.043 -19414.132
PROX_CBD -76833.361 5767.956 -0.262 -13.321 0.000 -88147.991 -65518.730
PROX_CHILDCARE -297608.214 109400.497 -0.078 -2.720 0.007 -512212.168 -83004.261
PROX_ELDERLYCARE 183303.549 39943.561 0.089 4.589 0.000 104948.823 261658.276
PROX_URA_GROWTH_AREA 39752.039 11763.983 0.061 3.379 0.001 16675.385 62828.692
PROX_MRT -305114.878 57591.189 -0.116 -5.298 0.000 -418087.828 -192141.927
PROX_PARK 572038.799 65511.407 0.150 8.732 0.000 443529.265 700548.334
PROX_PRIMARY_SCH 164542.899 60358.977 0.064 2.726 0.006 46140.557 282945.241
PROX_SHOPPING_MALL -220515.279 36558.846 -0.115 -6.032 0.000 -292230.427 -148800.131
PROX_BUS_STOP 674997.951 134646.651 0.133 5.013 0.000 410870.234 939125.668
NO_Of_UNITS -228.616 89.102 -0.049 -2.566 0.010 -403.402 -53.830
FAMILY_FRIENDLY 148152.863 46913.189 0.058 3.158 0.002 56126.263 240179.463
FREEHOLD 281136.713 76537.974 0.109 3.673 0.000 130997.067 431276.358
LEASEHOLD_99YR -89655.454 76421.659 -0.035 -1.173 0.241 -239566.931 60256.022
-----------------------------------------------------------------------------------------------------------------
We use the olsrrr package to generate a report for the model we created above (condo.mlr), similar to the method shown in the above code chunk.
ols_regress(condo.mlr) Model Summary
-----------------------------------------------------------------------------
R 0.807 RMSE 750537.537
R-Squared 0.652 MSE 571262902261.223
Adj. R-Squared 0.647 Coef. Var 43.160
Pred R-Squared 0.637 AIC 42971.173
MAE 412117.987 SBC 43081.835
-----------------------------------------------------------------------------
RMSE: Root Mean Square Error
MSE: Mean Square Error
MAE: Mean Absolute Error
AIC: Akaike Information Criteria
SBC: Schwarz Bayesian Criteria
ANOVA
--------------------------------------------------------------------------------
Sum of
Squares DF Mean Square F Sig.
--------------------------------------------------------------------------------
Regression 1.515738e+15 19 7.977571e+13 139.648 0.0000
Residual 8.089083e+14 1416 571262902261.223
Total 2.324647e+15 1435
--------------------------------------------------------------------------------
Parameter Estimates
-----------------------------------------------------------------------------------------------------------------
model Beta Std. Error Std. Beta t Sig lower upper
-----------------------------------------------------------------------------------------------------------------
(Intercept) 543071.420 136210.918 3.987 0.000 275874.535 810268.305
AREA_SQM 12688.669 370.119 0.579 34.283 0.000 11962.627 13414.710
AGE -24566.001 2766.041 -0.166 -8.881 0.000 -29991.980 -19140.022
PROX_CBD -78121.985 6791.377 -0.267 -11.503 0.000 -91444.227 -64799.744
PROX_CHILDCARE -333219.036 111020.303 -0.087 -3.001 0.003 -551000.984 -115437.089
PROX_ELDERLYCARE 170949.961 42110.748 0.083 4.060 0.000 88343.803 253556.120
PROX_URA_GROWTH_AREA 38507.622 12523.661 0.059 3.075 0.002 13940.700 63074.545
PROX_HAWKER_MARKET 23801.197 29299.923 0.019 0.812 0.417 -33674.725 81277.120
PROX_KINDERGARTEN 144097.972 82738.669 0.030 1.742 0.082 -18205.570 306401.514
PROX_MRT -322775.874 58528.079 -0.123 -5.515 0.000 -437586.937 -207964.811
PROX_PARK 564487.876 66563.011 0.148 8.481 0.000 433915.162 695060.590
PROX_PRIMARY_SCH 186170.524 65515.193 0.072 2.842 0.005 57653.253 314687.795
PROX_TOP_PRIMARY_SCH -477.073 20597.972 -0.001 -0.023 0.982 -40882.894 39928.747
PROX_SHOPPING_MALL -207721.520 42855.500 -0.109 -4.847 0.000 -291788.613 -123654.427
PROX_SUPERMARKET -48074.679 77145.257 -0.012 -0.623 0.533 -199405.956 103256.599
PROX_BUS_STOP 675755.044 138551.991 0.133 4.877 0.000 403965.817 947544.272
NO_Of_UNITS -216.180 90.302 -0.046 -2.394 0.017 -393.320 -39.040
FAMILY_FRIENDLY 142128.272 47055.082 0.056 3.020 0.003 49823.107 234433.438
FREEHOLD 300646.543 77296.529 0.117 3.890 0.000 149018.525 452274.561
LEASEHOLD_99YR -77137.375 77570.869 -0.030 -0.994 0.320 -229303.551 75028.801
-----------------------------------------------------------------------------------------------------------------
This method even shows you the meaning of RMSE and other important statistics and facilitates analysis further.
It is a better method to use for detailed output compared to the basic lm() method.
Adjusted R-squared value tells you how good your model is by informing you how much of the variation in price our model accounts for.
Multicollinearity check using vif
We can check for multicollinearity using the code chunk below. (vif= variance inflation factor)
ols_vif_tol(condo.mlr) Variables Tolerance VIF
1 AREA_SQM 0.8601326 1.162611
2 AGE 0.7011585 1.426211
3 PROX_CBD 0.4575471 2.185567
4 PROX_CHILDCARE 0.2898233 3.450378
5 PROX_ELDERLYCARE 0.5922238 1.688551
6 PROX_URA_GROWTH_AREA 0.6614081 1.511926
7 PROX_HAWKER_MARKET 0.4373874 2.286303
8 PROX_KINDERGARTEN 0.8356793 1.196631
9 PROX_MRT 0.4949877 2.020252
10 PROX_PARK 0.8015728 1.247547
11 PROX_PRIMARY_SCH 0.3823248 2.615577
12 PROX_TOP_PRIMARY_SCH 0.4878620 2.049760
13 PROX_SHOPPING_MALL 0.4903052 2.039546
14 PROX_SUPERMARKET 0.6142127 1.628100
15 PROX_BUS_STOP 0.3311024 3.020213
16 NO_Of_UNITS 0.6543336 1.528272
17 FAMILY_FRIENDLY 0.7191719 1.390488
18 FREEHOLD 0.2728521 3.664990
19 LEASEHOLD_99YR 0.2645988 3.779307
Higher vif indicates (over 5) multicollinearity. Above 10 means they must be eliminated from the model.
In our case, it shows that we do not need to eliminate them.
Highly continuous correlated variables impact the model more so than correlated dummy variables would.
Variable Selection using Stepwise Regression
There is forward and backward stepwise regression.
Forward: Start with no variables and add them one by one based on model improvement.
Backward: Start with all variables and remove them one by one based on model deterioration.
condo_fw_mlr=ols_step_forward_p(condo.mlr,
p_val = 0.05,
details = FALSE)
condo_fw_mlr
Stepwise Summary
-----------------------------------------------------------------------------------------
Step Variable AIC SBC SBIC R2 Adj. R2
-----------------------------------------------------------------------------------------
0 Base Model 44449.068 44459.608 40371.745 0.00000 0.00000
1 AREA_SQM 43587.753 43603.562 39510.883 0.45184 0.45146
2 PROX_CBD 43243.523 43264.602 39167.182 0.56928 0.56868
3 PROX_PARK 43177.691 43204.039 39101.331 0.58915 0.58829
4 FREEHOLD 43125.474 43157.092 39049.179 0.60438 0.60327
5 AGE 43069.222 43106.109 38993.167 0.62010 0.61878
6 PROX_ELDERLYCARE 43046.515 43088.672 38970.548 0.62659 0.62502
7 PROX_SHOPPING_MALL 43020.990 43068.417 38945.209 0.63367 0.63188
8 PROX_URA_GROWTH_AREA 43009.092 43061.788 38933.407 0.63720 0.63517
9 PROX_MRT 42999.058 43057.024 38923.483 0.64023 0.63796
10 PROX_BUS_STOP 42984.951 43048.186 38909.582 0.64424 0.64175
11 FAMILY_FRIENDLY 42981.085 43049.590 38905.797 0.64569 0.64296
12 NO_Of_UNITS 42975.246 43049.021 38900.092 0.64762 0.64465
13 PROX_CHILDCARE 42971.858 43050.902 38896.812 0.64894 0.64573
14 PROX_PRIMARY_SCH 42966.758 43051.072 38891.872 0.65067 0.64723
-----------------------------------------------------------------------------------------
Final Model Output
------------------
Model Summary
-----------------------------------------------------------------------------
R 0.807 RMSE 751998.679
R-Squared 0.651 MSE 571471422208.591
Adj. R-Squared 0.647 Coef. Var 43.168
Pred R-Squared 0.638 AIC 42966.758
MAE 414819.628 SBC 43051.072
-----------------------------------------------------------------------------
RMSE: Root Mean Square Error
MSE: Mean Square Error
MAE: Mean Absolute Error
AIC: Akaike Information Criteria
SBC: Schwarz Bayesian Criteria
ANOVA
--------------------------------------------------------------------------------
Sum of
Squares DF Mean Square F Sig.
--------------------------------------------------------------------------------
Regression 1.512586e+15 14 1.080418e+14 189.059 0.0000
Residual 8.120609e+14 1421 571471422208.591
Total 2.324647e+15 1435
--------------------------------------------------------------------------------
Parameter Estimates
-----------------------------------------------------------------------------------------------------------------
model Beta Std. Error Std. Beta t Sig lower upper
-----------------------------------------------------------------------------------------------------------------
(Intercept) 527633.222 108183.223 4.877 0.000 315417.244 739849.200
AREA_SQM 12777.523 367.479 0.584 34.771 0.000 12056.663 13498.382
PROX_CBD -77131.323 5763.125 -0.263 -13.384 0.000 -88436.469 -65826.176
PROX_PARK 570504.807 65507.029 0.150 8.709 0.000 442003.938 699005.677
FREEHOLD 350599.812 48506.485 0.136 7.228 0.000 255447.802 445751.821
AGE -24687.739 2754.845 -0.167 -8.962 0.000 -30091.739 -19283.740
PROX_ELDERLYCARE 185575.623 39901.864 0.090 4.651 0.000 107302.737 263848.510
PROX_SHOPPING_MALL -220947.251 36561.832 -0.115 -6.043 0.000 -292668.213 -149226.288
PROX_URA_GROWTH_AREA 39163.254 11754.829 0.060 3.332 0.001 16104.571 62221.936
PROX_MRT -294745.107 56916.367 -0.112 -5.179 0.000 -406394.234 -183095.980
PROX_BUS_STOP 682482.221 134513.243 0.134 5.074 0.000 418616.359 946348.082
FAMILY_FRIENDLY 146307.576 46893.021 0.057 3.120 0.002 54320.593 238294.560
NO_Of_UNITS -245.480 87.947 -0.053 -2.791 0.005 -418.000 -72.961
PROX_CHILDCARE -318472.751 107959.512 -0.084 -2.950 0.003 -530249.889 -106695.613
PROX_PRIMARY_SCH 159856.136 60234.599 0.062 2.654 0.008 41697.849 278014.424
-----------------------------------------------------------------------------------------------------------------
plot(condo_fw_mlr)
condo_bw_mlr=ols_step_backward_p(condo.mlr,
p_val = 0.05,
details = FALSE)
condo_bw_mlr
Stepwise Summary
-----------------------------------------------------------------------------------------
Step Variable AIC SBC SBIC R2 Adj. R2
-----------------------------------------------------------------------------------------
0 Full Model 42971.173 43081.835 38896.546 0.65203 0.64736
1 PROX_TOP_PRIMARY_SCH 42969.173 43074.565 38894.518 0.65203 0.64761
2 PROX_SUPERMARKET 42967.567 43067.689 38892.873 0.65193 0.64776
3 PROX_HAWKER_MARKET 42966.461 43061.315 38891.719 0.65172 0.64779
4 LEASEHOLD_99YR 42965.558 43055.141 38890.764 0.65145 0.64777
5 PROX_KINDERGARTEN 42966.758 43051.072 38891.872 0.65067 0.64723
-----------------------------------------------------------------------------------------
Final Model Output
------------------
Model Summary
-----------------------------------------------------------------------------
R 0.807 RMSE 751998.679
R-Squared 0.651 MSE 571471422208.591
Adj. R-Squared 0.647 Coef. Var 43.168
Pred R-Squared 0.638 AIC 42966.758
MAE 414819.628 SBC 43051.072
-----------------------------------------------------------------------------
RMSE: Root Mean Square Error
MSE: Mean Square Error
MAE: Mean Absolute Error
AIC: Akaike Information Criteria
SBC: Schwarz Bayesian Criteria
ANOVA
--------------------------------------------------------------------------------
Sum of
Squares DF Mean Square F Sig.
--------------------------------------------------------------------------------
Regression 1.512586e+15 14 1.080418e+14 189.059 0.0000
Residual 8.120609e+14 1421 571471422208.591
Total 2.324647e+15 1435
--------------------------------------------------------------------------------
Parameter Estimates
-----------------------------------------------------------------------------------------------------------------
model Beta Std. Error Std. Beta t Sig lower upper
-----------------------------------------------------------------------------------------------------------------
(Intercept) 527633.222 108183.223 4.877 0.000 315417.244 739849.200
AREA_SQM 12777.523 367.479 0.584 34.771 0.000 12056.663 13498.382
AGE -24687.739 2754.845 -0.167 -8.962 0.000 -30091.739 -19283.740
PROX_CBD -77131.323 5763.125 -0.263 -13.384 0.000 -88436.469 -65826.176
PROX_CHILDCARE -318472.751 107959.512 -0.084 -2.950 0.003 -530249.889 -106695.613
PROX_ELDERLYCARE 185575.623 39901.864 0.090 4.651 0.000 107302.737 263848.510
PROX_URA_GROWTH_AREA 39163.254 11754.829 0.060 3.332 0.001 16104.571 62221.936
PROX_MRT -294745.107 56916.367 -0.112 -5.179 0.000 -406394.234 -183095.980
PROX_PARK 570504.807 65507.029 0.150 8.709 0.000 442003.938 699005.677
PROX_PRIMARY_SCH 159856.136 60234.599 0.062 2.654 0.008 41697.849 278014.424
PROX_SHOPPING_MALL -220947.251 36561.832 -0.115 -6.043 0.000 -292668.213 -149226.288
PROX_BUS_STOP 682482.221 134513.243 0.134 5.074 0.000 418616.359 946348.082
NO_Of_UNITS -245.480 87.947 -0.053 -2.791 0.005 -418.000 -72.961
FAMILY_FRIENDLY 146307.576 46893.021 0.057 3.120 0.002 54320.593 238294.560
FREEHOLD 350599.812 48506.485 0.136 7.228 0.000 255447.802 445751.821
-----------------------------------------------------------------------------------------------------------------
plot(condo_bw_mlr)
7.4.5 Preparing Publication Quality Table: gtsummary method
The broom package provides an elegant and flexible way to create publication-ready summary tables in R.
In the code chunk below, tidy() function is used to create a well formatted regression report.
broom::tidy(condo.mlr1, intercept = TRUE)# A tibble: 16 × 5
term estimate std.error statistic p.value
<chr> <dbl> <dbl> <dbl> <dbl>
1 (Intercept) 591540. 121111. 4.88 1.16e- 6
2 AREA_SQM 12754. 368. 34.7 2.84e-191
3 AGE -24822. 2757. -9.00 6.84e- 19
4 PROX_CBD -76833. 5768. -13.3 3.11e- 38
5 PROX_CHILDCARE -297608. 109400. -2.72 6.60e- 3
6 PROX_ELDERLYCARE 183304. 39944. 4.59 4.85e- 6
7 PROX_URA_GROWTH_AREA 39752. 11764. 3.38 7.47e- 4
8 PROX_MRT -305115. 57591. -5.30 1.36e- 7
9 PROX_PARK 572039. 65511. 8.73 6.91e- 18
10 PROX_PRIMARY_SCH 164543. 60359. 2.73 6.49e- 3
11 PROX_SHOPPING_MALL -220515. 36559. -6.03 2.06e- 9
12 PROX_BUS_STOP 674998. 134647. 5.01 6.03e- 7
13 NO_Of_UNITS -229. 89.1 -2.57 1.04e- 2
14 FAMILY_FRIENDLY 148153. 46913. 3.16 1.62e- 3
15 FREEHOLD 281137. 76538. 3.67 2.48e- 4
16 LEASEHOLD_99YR -89655. 76422. -1.17 2.41e- 1
7.4.6 Checking for Multicollinearity
In this section, we use anl R package designed specifically for conducting OLS (Ordinary Least Squares) regression analysis—olsrr. This package offers a wide range of valuable tools to help you build more robust multiple linear regression models. Its key features include:
- Comprehensive regression output
- Diagnostic tests for residuals
- Influence measures for identifying outliers
- Tests for heteroskedasticity
- Collinearity diagnostics to detect multicollinearity
- Model fit assessment
- Evaluation of variable contributions
- Various methods for variable selection
In the code snippet below, we demonstrate how to use the ols_vif_tol() function from the olsrr package to assess potential multicollinearity among predictors in your regression model.
ols_vif_tol(condo.mlr1) Variables Tolerance VIF
1 AREA_SQM 0.8703348 1.148983
2 AGE 0.7059074 1.416616
3 PROX_CBD 0.6343823 1.576337
4 PROX_CHILDCARE 0.2984991 3.350094
5 PROX_ELDERLYCARE 0.6582967 1.519072
6 PROX_URA_GROWTH_AREA 0.7496642 1.333931
7 PROX_MRT 0.5112747 1.955896
8 PROX_PARK 0.8275963 1.208319
9 PROX_PRIMARY_SCH 0.4504807 2.219851
10 PROX_SHOPPING_MALL 0.6738111 1.484095
11 PROX_BUS_STOP 0.3506229 2.852067
12 NO_Of_UNITS 0.6721417 1.487781
13 FAMILY_FRIENDLY 0.7236014 1.381976
14 FREEHOLD 0.2783151 3.593049
15 LEASEHOLD_99YR 0.2726438 3.667789
Since the VIF of the independent variables are less than 10. We can safely conclude that there are no sign of multicollinearity among the independent variables.
7.4.6.1 Test for Non-Linearity
In multiple linear regression, it is important for us to test the assumption that linearity and additivity of the relationship between dependent and independent variables.
In the code chunk below, the ols_plot_resid_fit() of olsrr package is used to perform linearity assumption test.
Notice that we pass the newly generated forward stepwise model we created above. We explicitly state that it must use the model.
ols_plot_resid_fit(condo_fw_mlr$model)
The figure above reveals that most of the data points are scattered around the 0 line, hence we can safely conclude that the relationships between the dependent variable and independent variables are linear.
7.4.6.2 Test for Normality Assumption
The code chunk below uses ols_plot_resid_hist() of olsrr package to perform normality assumption test.
ols_plot_resid_hist(condo_fw_mlr$model)
The figure reveals that the residual of the multiple linear regression model (i.e. condo.mlr1) is resemble normal distribution.
If you prefer formal statistical test methods, the ols_test_normality() of olsrr package can be used as shown in the code chun below.
ols_test_normality(condo_fw_mlr$model)-----------------------------------------------
Test Statistic pvalue
-----------------------------------------------
Shapiro-Wilk 0.6856 0.0000
Kolmogorov-Smirnov 0.1366 0.0000
Cramer-von Mises 121.0768 0.0000
Anderson-Darling 67.9551 0.0000
-----------------------------------------------
The summary table above reveals that the p-values of the four tests are way smaller than the alpha value of 0.05. Hence we will reject the null hypothesis and infer that there is statistical evidence that the residual are not normally distributed.
7.4.6.3 Testing for Spatial Autocorrelation
The hedonic model we try to build are using geographically referenced attributes, hence it is also important for us to visual the residual of the hedonic pricing model.
In order to perform spatial autocorrelation test, we need to convert condo_resale.sf from sf data frame into a SpatialPointsDataFrame.
First, we will export the residual of the hedonic pricing model and save it as a data frame.
mlr.output <- as.data.frame(condo_fw_mlr$model$residuals)%>%
rename(`FW_MLR_RES` = `condo_fw_mlr$model$residuals`)Next, we will join the newly created data frame with condo_resale.sf object.
condo_resale.sf <- cbind(condo_resale.sf,
mlr.output$`FW_MLR_RES`)%>%
rename('MLR_RES'='mlr.output.FW_MLR_RES')Next, we will convert condo_resale.res.sf from simple feature object into a SpatialPointsDataFrame because spdep package can only process sp conformed spatial data objects.
Next, we will use tmap package to display the distribution of the residuals on an interactive map.
tmap_mode("view")
tm_shape(mpsz_svy21)+
tmap_options(check.and.fix = TRUE) + # If we know that this particular layer is cause us issues, we specifically mention check.and.fix=TRUE here.
tm_polygons(alpha = 0.4) +
tm_shape(condo_resale.sf) +
tm_dots(col = "MLR_RES",
alpha = 0.6,
style="quantile") +
tm_view(set.zoom.limits = c(11,14))tmap_mode('plot')Residuals show the difference between actual transaction price and estimated price by the model.
TO prove that our observation is indeed true, the Moran’s I test will be performed.
We first compute the distance-based weight matrix by using the st_knn() function of the sfdep package.
condo_resale.sf= condo_resale.sf%>%
mutate(nb=st_knn(geometry, k=6,
longlat=FALSE),
wt= st_weights(nb,
style='W'),
.before=1)global_moran_perm(condo_resale.sf$MLR_RES,
condo_resale.sf$nb,
condo_resale.sf$wt,
alternative='two.sided',
nsim=99)
Monte-Carlo simulation of Moran I
data: x
weights: listw
number of simulations + 1: 100
statistic = 0.32254, observed rank = 100, p-value < 2.2e-16
alternative hypothesis: two.sided
The global Moran’s I test for residual spatial autocorrelation shows that it’s P value is less than 0.05, meaning we have sufficient evidence to reject the null hypothesis that the residuals are randomly distributed.
Since the observed Global Moran I=0.25586, which is greater than 0, we can infer that the residuals resemble cluster distribution.
7.5 Building Hedonic Pricing Models using GWModel
7.5.1 Building Fixed Bandwidth GWR Model
7.5.1.1 Computing fixed bandwith
In the code chunk below, the bw.gwr() function of the GWModel package is used to determine the optimal fixed bandwidth to use in the model.
Notice that the argument adaptive is set to FALSE indicates that we are interested to compute the fixed bandwidth.
There are two possible approaches can be used to determine the stopping rule.
CV cross-validation approach
AIC corrected (AICc) approach.
We define the stopping rule using approach argeement.
bw.fixed <- bw.gwr(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP + NO_Of_UNITS +
FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale.sf,
approach="CV",
kernel="gaussian",
adaptive=FALSE,
longlat=FALSE)Fixed bandwidth: 17660.96 CV score: 8.259118e+14
Fixed bandwidth: 10917.26 CV score: 7.970454e+14
Fixed bandwidth: 6749.419 CV score: 7.273273e+14
Fixed bandwidth: 4173.553 CV score: 6.300006e+14
Fixed bandwidth: 2581.58 CV score: 5.404958e+14
Fixed bandwidth: 1597.687 CV score: 4.857515e+14
Fixed bandwidth: 989.6077 CV score: 4.722431e+14
Fixed bandwidth: 613.7939 CV score: 1.378294e+16
Fixed bandwidth: 1221.873 CV score: 4.778717e+14
Fixed bandwidth: 846.0596 CV score: 4.791629e+14
Fixed bandwidth: 1078.325 CV score: 4.751406e+14
Fixed bandwidth: 934.7772 CV score: 4.72518e+14
Fixed bandwidth: 1023.495 CV score: 4.730305e+14
Fixed bandwidth: 968.6643 CV score: 4.721317e+14
Fixed bandwidth: 955.7206 CV score: 4.722072e+14
Fixed bandwidth: 976.6639 CV score: 4.721387e+14
Fixed bandwidth: 963.7202 CV score: 4.721484e+14
Fixed bandwidth: 971.7199 CV score: 4.721293e+14
Fixed bandwidth: 973.6083 CV score: 4.721309e+14
Fixed bandwidth: 970.5527 CV score: 4.721295e+14
Fixed bandwidth: 972.4412 CV score: 4.721296e+14
Fixed bandwidth: 971.2741 CV score: 4.721292e+14
Fixed bandwidth: 970.9985 CV score: 4.721293e+14
Fixed bandwidth: 971.4443 CV score: 4.721292e+14
Fixed bandwidth: 971.5496 CV score: 4.721293e+14
Fixed bandwidth: 971.3793 CV score: 4.721292e+14
Fixed bandwidth: 971.3391 CV score: 4.721292e+14
Fixed bandwidth: 971.3143 CV score: 4.721292e+14
Fixed bandwidth: 971.3545 CV score: 4.721292e+14
Fixed bandwidth: 971.3296 CV score: 4.721292e+14
Fixed bandwidth: 971.345 CV score: 4.721292e+14
Fixed bandwidth: 971.3355 CV score: 4.721292e+14
Fixed bandwidth: 971.3413 CV score: 4.721292e+14
Fixed bandwidth: 971.3377 CV score: 4.721292e+14
Fixed bandwidth: 971.34 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
Fixed bandwidth: 971.3408 CV score: 4.721292e+14
Fixed bandwidth: 971.3403 CV score: 4.721292e+14
Fixed bandwidth: 971.3406 CV score: 4.721292e+14
Fixed bandwidth: 971.3404 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
The result shows that the recommended bandwidth is 971.3405 metres.
7.5.1.2 GWModel method - fixed bandwith
Now we can use the code chunk below to calibrate the gwr model using fixed bandwidth and gaussian kernel.
gwr.fixed <- bw.gwr(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH +
PROX_SHOPPING_MALL + PROX_BUS_STOP + NO_Of_UNITS +
FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale.sf,
approach='CV',
kernel = 'gaussian',
adaptive = FALSE,
longlat = FALSE)Fixed bandwidth: 17660.96 CV score: 8.259118e+14
Fixed bandwidth: 10917.26 CV score: 7.970454e+14
Fixed bandwidth: 6749.419 CV score: 7.273273e+14
Fixed bandwidth: 4173.553 CV score: 6.300006e+14
Fixed bandwidth: 2581.58 CV score: 5.404958e+14
Fixed bandwidth: 1597.687 CV score: 4.857515e+14
Fixed bandwidth: 989.6077 CV score: 4.722431e+14
Fixed bandwidth: 613.7939 CV score: 1.378294e+16
Fixed bandwidth: 1221.873 CV score: 4.778717e+14
Fixed bandwidth: 846.0596 CV score: 4.791629e+14
Fixed bandwidth: 1078.325 CV score: 4.751406e+14
Fixed bandwidth: 934.7772 CV score: 4.72518e+14
Fixed bandwidth: 1023.495 CV score: 4.730305e+14
Fixed bandwidth: 968.6643 CV score: 4.721317e+14
Fixed bandwidth: 955.7206 CV score: 4.722072e+14
Fixed bandwidth: 976.6639 CV score: 4.721387e+14
Fixed bandwidth: 963.7202 CV score: 4.721484e+14
Fixed bandwidth: 971.7199 CV score: 4.721293e+14
Fixed bandwidth: 973.6083 CV score: 4.721309e+14
Fixed bandwidth: 970.5527 CV score: 4.721295e+14
Fixed bandwidth: 972.4412 CV score: 4.721296e+14
Fixed bandwidth: 971.2741 CV score: 4.721292e+14
Fixed bandwidth: 970.9985 CV score: 4.721293e+14
Fixed bandwidth: 971.4443 CV score: 4.721292e+14
Fixed bandwidth: 971.5496 CV score: 4.721293e+14
Fixed bandwidth: 971.3793 CV score: 4.721292e+14
Fixed bandwidth: 971.3391 CV score: 4.721292e+14
Fixed bandwidth: 971.3143 CV score: 4.721292e+14
Fixed bandwidth: 971.3545 CV score: 4.721292e+14
Fixed bandwidth: 971.3296 CV score: 4.721292e+14
Fixed bandwidth: 971.345 CV score: 4.721292e+14
Fixed bandwidth: 971.3355 CV score: 4.721292e+14
Fixed bandwidth: 971.3413 CV score: 4.721292e+14
Fixed bandwidth: 971.3377 CV score: 4.721292e+14
Fixed bandwidth: 971.34 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
Fixed bandwidth: 971.3408 CV score: 4.721292e+14
Fixed bandwidth: 971.3403 CV score: 4.721292e+14
Fixed bandwidth: 971.3406 CV score: 4.721292e+14
Fixed bandwidth: 971.3404 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
Fixed bandwidth: 971.3405 CV score: 4.721292e+14
We use the code chunk below to display the model created above.
gwr.fixed[1] 971.3405
gwr.fixed<- gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_MRT + PROX_PARK +
PROX_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale.sf,
bw=bw.fixed,
kernel = 'gaussian',
longlat = FALSE)The output is saved in a list of flass ‘gwrm’. The code chunk below is used to display the model output.
gwr.fixed ***********************************************************************
* Package GWmodel *
***********************************************************************
Program starts at: 2024-10-17 13:24:34.435927
Call:
gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH + PROX_SHOPPING_MALL +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data = condo_resale.sf, bw = bw.fixed, kernel = "gaussian",
longlat = FALSE)
Dependent (y) variable: SELLING_PRICE
Independent variables: AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE PROX_URA_GROWTH_AREA PROX_MRT PROX_PARK PROX_PRIMARY_SCH PROX_SHOPPING_MALL PROX_BUS_STOP NO_Of_UNITS FAMILY_FRIENDLY FREEHOLD
Number of data points: 1436
***********************************************************************
* Results of Global Regression *
***********************************************************************
Call:
lm(formula = formula, data = data)
Residuals:
Min 1Q Median 3Q Max
-3470778 -298119 -23481 248917 12234210
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 527633.22 108183.22 4.877 1.20e-06 ***
AREA_SQM 12777.52 367.48 34.771 < 2e-16 ***
AGE -24687.74 2754.84 -8.962 < 2e-16 ***
PROX_CBD -77131.32 5763.12 -13.384 < 2e-16 ***
PROX_CHILDCARE -318472.75 107959.51 -2.950 0.003231 **
PROX_ELDERLYCARE 185575.62 39901.86 4.651 3.61e-06 ***
PROX_URA_GROWTH_AREA 39163.25 11754.83 3.332 0.000885 ***
PROX_MRT -294745.11 56916.37 -5.179 2.56e-07 ***
PROX_PARK 570504.81 65507.03 8.709 < 2e-16 ***
PROX_PRIMARY_SCH 159856.14 60234.60 2.654 0.008046 **
PROX_SHOPPING_MALL -220947.25 36561.83 -6.043 1.93e-09 ***
PROX_BUS_STOP 682482.22 134513.24 5.074 4.42e-07 ***
NO_Of_UNITS -245.48 87.95 -2.791 0.005321 **
FAMILY_FRIENDLY 146307.58 46893.02 3.120 0.001845 **
FREEHOLD 350599.81 48506.48 7.228 7.98e-13 ***
---Significance stars
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 756000 on 1421 degrees of freedom
Multiple R-squared: 0.6507
Adjusted R-squared: 0.6472
F-statistic: 189.1 on 14 and 1421 DF, p-value: < 2.2e-16
***Extra Diagnostic information
Residual sum of squares: 8.120609e+14
Sigma(hat): 752522.9
AIC: 42966.76
AICc: 42967.14
BIC: 41731.39
***********************************************************************
* Results of Geographically Weighted Regression *
***********************************************************************
*********************Model calibration information*********************
Kernel function: gaussian
Fixed bandwidth: 971.3405
Regression points: the same locations as observations are used.
Distance metric: Euclidean distance metric is used.
****************Summary of GWR coefficient estimates:******************
Min. 1st Qu. Median 3rd Qu.
Intercept -3.5988e+07 -5.1998e+05 7.6780e+05 1.7412e+06
AREA_SQM 1.0003e+03 5.2758e+03 7.4740e+03 1.2301e+04
AGE -1.3475e+05 -2.0813e+04 -8.6260e+03 -3.7784e+03
PROX_CBD -7.7047e+07 -2.3608e+05 -8.3600e+04 3.4646e+04
PROX_CHILDCARE -6.0097e+06 -3.3667e+05 -9.7425e+04 2.9007e+05
PROX_ELDERLYCARE -3.5000e+06 -1.5970e+05 3.1971e+04 1.9577e+05
PROX_URA_GROWTH_AREA -3.0170e+06 -8.2013e+04 7.0749e+04 2.2612e+05
PROX_MRT -3.5282e+06 -6.5836e+05 -1.8833e+05 3.6922e+04
PROX_PARK -1.2062e+06 -2.1732e+05 3.5383e+04 4.1335e+05
PROX_PRIMARY_SCH -2.2695e+07 -1.7066e+05 4.8472e+04 5.1555e+05
PROX_SHOPPING_MALL -7.2585e+06 -1.6684e+05 -1.0517e+04 1.5923e+05
PROX_BUS_STOP -1.4676e+06 -4.5207e+04 3.7601e+05 1.1664e+06
NO_Of_UNITS -1.3170e+03 -2.4822e+02 -3.0846e+01 2.5496e+02
FAMILY_FRIENDLY -2.2749e+06 -1.1140e+05 7.6214e+03 1.6107e+05
FREEHOLD -9.2067e+06 3.8073e+04 1.5169e+05 3.7528e+05
Max.
Intercept 112793548
AREA_SQM 21575
AGE 434201
PROX_CBD 2704596
PROX_CHILDCARE 1654087
PROX_ELDERLYCARE 38867814
PROX_URA_GROWTH_AREA 78515730
PROX_MRT 3124316
PROX_PARK 18122425
PROX_PRIMARY_SCH 4637503
PROX_SHOPPING_MALL 1529952
PROX_BUS_STOP 11342182
NO_Of_UNITS 12907
FAMILY_FRIENDLY 1720744
FREEHOLD 6073636
************************Diagnostic information*************************
Number of data points: 1436
Effective number of parameters (2trace(S) - trace(S'S)): 438.3804
Effective degrees of freedom (n-2trace(S) + trace(S'S)): 997.6196
AICc (GWR book, Fotheringham, et al. 2002, p. 61, eq 2.33): 42263.61
AIC (GWR book, Fotheringham, et al. 2002,GWR p. 96, eq. 4.22): 41632.36
BIC (GWR book, Fotheringham, et al. 2002,GWR p. 61, eq. 2.34): 42515.71
Residual sum of squares: 2.53407e+14
R-square value: 0.8909912
Adjusted R-square value: 0.8430417
***********************************************************************
Program stops at: 2024-10-17 13:24:36.082047
Notice the change, the improvement, in Adjusted R-Squared Value.
7.5.2 Building Adaptive Bandwidth GWR Model
7.5.2.1 Computing the adaptive bandwidth
Similar to the earlier section, we will first use bw.gwr() to determine the recommended data point to use.
The code chunk used look very similar to the one used to compute the fixed bandwidth except the adaptive argument has changed to TRUE.
bw.adaptive <- bw.gwr(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_MRT + PROX_PARK +
PROX_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale.sf,
approach="CV",
kernel="gaussian",
adaptive=TRUE,
longlat=FALSE)Adaptive bandwidth: 895 CV score: 7.952401e+14
Adaptive bandwidth: 561 CV score: 7.667364e+14
Adaptive bandwidth: 354 CV score: 6.953454e+14
Adaptive bandwidth: 226 CV score: 6.15223e+14
Adaptive bandwidth: 147 CV score: 5.674373e+14
Adaptive bandwidth: 98 CV score: 5.426745e+14
Adaptive bandwidth: 68 CV score: 5.168117e+14
Adaptive bandwidth: 49 CV score: 4.859631e+14
Adaptive bandwidth: 37 CV score: 4.646518e+14
Adaptive bandwidth: 30 CV score: 4.422088e+14
Adaptive bandwidth: 25 CV score: 4.430816e+14
Adaptive bandwidth: 32 CV score: 4.505602e+14
Adaptive bandwidth: 27 CV score: 4.462172e+14
Adaptive bandwidth: 30 CV score: 4.422088e+14
Constructing Adaptive Model
gwr.adaptive <- gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE +
PROX_CBD + PROX_CHILDCARE + PROX_ELDERLYCARE +
PROX_URA_GROWTH_AREA + PROX_MRT + PROX_PARK +
PROX_PRIMARY_SCH + PROX_SHOPPING_MALL + PROX_BUS_STOP +
NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data=condo_resale.sf, bw=bw.adaptive,
kernel = 'gaussian',
adaptive=TRUE,
longlat = FALSE)The code chunk below can be used to display the model output.
gwr.adaptive ***********************************************************************
* Package GWmodel *
***********************************************************************
Program starts at: 2024-10-17 13:24:46.474889
Call:
gwr.basic(formula = SELLING_PRICE ~ AREA_SQM + AGE + PROX_CBD +
PROX_CHILDCARE + PROX_ELDERLYCARE + PROX_URA_GROWTH_AREA +
PROX_MRT + PROX_PARK + PROX_PRIMARY_SCH + PROX_SHOPPING_MALL +
PROX_BUS_STOP + NO_Of_UNITS + FAMILY_FRIENDLY + FREEHOLD,
data = condo_resale.sf, bw = bw.adaptive, kernel = "gaussian",
adaptive = TRUE, longlat = FALSE)
Dependent (y) variable: SELLING_PRICE
Independent variables: AREA_SQM AGE PROX_CBD PROX_CHILDCARE PROX_ELDERLYCARE PROX_URA_GROWTH_AREA PROX_MRT PROX_PARK PROX_PRIMARY_SCH PROX_SHOPPING_MALL PROX_BUS_STOP NO_Of_UNITS FAMILY_FRIENDLY FREEHOLD
Number of data points: 1436
***********************************************************************
* Results of Global Regression *
***********************************************************************
Call:
lm(formula = formula, data = data)
Residuals:
Min 1Q Median 3Q Max
-3470778 -298119 -23481 248917 12234210
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 527633.22 108183.22 4.877 1.20e-06 ***
AREA_SQM 12777.52 367.48 34.771 < 2e-16 ***
AGE -24687.74 2754.84 -8.962 < 2e-16 ***
PROX_CBD -77131.32 5763.12 -13.384 < 2e-16 ***
PROX_CHILDCARE -318472.75 107959.51 -2.950 0.003231 **
PROX_ELDERLYCARE 185575.62 39901.86 4.651 3.61e-06 ***
PROX_URA_GROWTH_AREA 39163.25 11754.83 3.332 0.000885 ***
PROX_MRT -294745.11 56916.37 -5.179 2.56e-07 ***
PROX_PARK 570504.81 65507.03 8.709 < 2e-16 ***
PROX_PRIMARY_SCH 159856.14 60234.60 2.654 0.008046 **
PROX_SHOPPING_MALL -220947.25 36561.83 -6.043 1.93e-09 ***
PROX_BUS_STOP 682482.22 134513.24 5.074 4.42e-07 ***
NO_Of_UNITS -245.48 87.95 -2.791 0.005321 **
FAMILY_FRIENDLY 146307.58 46893.02 3.120 0.001845 **
FREEHOLD 350599.81 48506.48 7.228 7.98e-13 ***
---Significance stars
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 756000 on 1421 degrees of freedom
Multiple R-squared: 0.6507
Adjusted R-squared: 0.6472
F-statistic: 189.1 on 14 and 1421 DF, p-value: < 2.2e-16
***Extra Diagnostic information
Residual sum of squares: 8.120609e+14
Sigma(hat): 752522.9
AIC: 42966.76
AICc: 42967.14
BIC: 41731.39
***********************************************************************
* Results of Geographically Weighted Regression *
***********************************************************************
*********************Model calibration information*********************
Kernel function: gaussian
Adaptive bandwidth: 30 (number of nearest neighbours)
Regression points: the same locations as observations are used.
Distance metric: Euclidean distance metric is used.
****************Summary of GWR coefficient estimates:******************
Min. 1st Qu. Median 3rd Qu.
Intercept -1.3487e+08 -2.4669e+05 7.7928e+05 1.6194e+06
AREA_SQM 3.3188e+03 5.6285e+03 7.7825e+03 1.2738e+04
AGE -9.6746e+04 -2.9288e+04 -1.4043e+04 -5.6119e+03
PROX_CBD -2.5330e+06 -1.6256e+05 -7.7242e+04 2.6624e+03
PROX_CHILDCARE -1.2790e+06 -2.0175e+05 8.7158e+03 3.7778e+05
PROX_ELDERLYCARE -1.6212e+06 -9.2050e+04 6.1029e+04 2.8184e+05
PROX_URA_GROWTH_AREA -7.2686e+06 -3.0350e+04 4.5869e+04 2.4613e+05
PROX_MRT -4.3781e+07 -6.7282e+05 -2.2115e+05 -7.4593e+04
PROX_PARK -2.9020e+06 -1.6782e+05 1.1601e+05 4.6572e+05
PROX_PRIMARY_SCH -8.6418e+05 -1.6627e+05 -7.7853e+03 4.3222e+05
PROX_SHOPPING_MALL -1.8272e+06 -1.3175e+05 -1.4049e+04 1.3799e+05
PROX_BUS_STOP -2.0579e+06 -7.1461e+04 4.1104e+05 1.2071e+06
NO_Of_UNITS -2.1993e+03 -2.3685e+02 -3.4699e+01 1.1657e+02
FAMILY_FRIENDLY -5.9879e+05 -5.0927e+04 2.6173e+04 2.2481e+05
FREEHOLD -1.6340e+05 4.0765e+04 1.9023e+05 3.7960e+05
Max.
Intercept 18758355
AREA_SQM 23064
AGE 13303
PROX_CBD 11346650
PROX_CHILDCARE 2892127
PROX_ELDERLYCARE 2465671
PROX_URA_GROWTH_AREA 7384059
PROX_MRT 1186242
PROX_PARK 2588497
PROX_PRIMARY_SCH 3381462
PROX_SHOPPING_MALL 38038564
PROX_BUS_STOP 12081592
NO_Of_UNITS 1010
FAMILY_FRIENDLY 2072414
FREEHOLD 1813995
************************Diagnostic information*************************
Number of data points: 1436
Effective number of parameters (2trace(S) - trace(S'S)): 350.3088
Effective degrees of freedom (n-2trace(S) + trace(S'S)): 1085.691
AICc (GWR book, Fotheringham, et al. 2002, p. 61, eq 2.33): 41982.22
AIC (GWR book, Fotheringham, et al. 2002,GWR p. 96, eq. 4.22): 41546.74
BIC (GWR book, Fotheringham, et al. 2002,GWR p. 61, eq. 2.34): 41914.08
Residual sum of squares: 2.528227e+14
R-square value: 0.8912425
Adjusted R-square value: 0.8561185
***********************************************************************
Program stops at: 2024-10-17 13:24:47.943393
7.5.3 Visualising GWR Output
In addition to the regression residuals, the output feature class table provides several key metrics, including observed and predicted values, the condition number (cond), Local R², residuals, and the coefficients with their standard errors for the explanatory variables:
Condition Number: This diagnostic assesses local collinearity. When strong local collinearity is present, model results become unstable. A condition number greater than 30 suggests that the results may be unreliable due to multicollinearity.
Local R²: Values range from 0.0 to 1.0 and indicate the goodness-of-fit of the local regression model. Low Local R² values signal poor model performance in those regions. Mapping these values can help identify areas where the Geographically Weighted Regression (GWR) model is performing well and where it is underperforming, potentially highlighting missing or unaccounted-for variables.
Predicted Values: These are the fitted y values estimated by the GWR model.
Residuals: Residuals are calculated by subtracting the fitted y values from the observed y values. Standardized residuals have a mean of zero and a standard deviation of one. A gradient map (cold-to-hot) of standardized residuals can be created to visualize areas of model under- or overestimation.
Coefficient Standard Errors: These values reflect the reliability of each coefficient estimate. Smaller standard errors relative to the actual coefficients indicate higher confidence in the estimates. Large standard errors, however, may suggest issues with local collinearity.
All of these metrics are stored within a SpatialPointsDataFrame or SpatialPolygonsDataFrame object, integrated with the fit points, GWR coefficient estimates, observed and predicted y values, coefficient standard errors, and t-values in the “data” slot of an object called SDF within the output list.
7.5.4 Converting SDF into sf data.frame
::: insights-box
SDF provides you data of the intercepts.
:::
::: {.cell}
```{.r .cell-code}
gwr.adaptive.output=as.data.frame(gwr.adaptive$SDF)%>%
select(-c(2:15)) # this removes the unnecessary columns and makes your work tidier.
:::
gwr_sf_adaptive=cbind(condo_resale.sf, gwr.adaptive.output)Next, glimpse() and summary() are used to display the content and summary of condo_resale.sf.adaptive sf data frame.
glimpse(gwr_sf_adaptive)Rows: 1,436
Columns: 64
$ nb <nb> <66, 77, 123, 238, 239, 343>, <21, 162, 163, 19…
$ wt <list> <0.1666667, 0.1666667, 0.1666667, 0.1666667, …
$ POSTCODE <dbl> 118635, 288420, 267833, 258380, 467169, 466472…
$ SELLING_PRICE <dbl> 3000000, 3880000, 3325000, 4250000, 1400000, 1…
$ AREA_SQM <dbl> 309, 290, 248, 127, 145, 139, 218, 141, 165, 1…
$ AGE <dbl> 30, 32, 33, 7, 28, 22, 24, 24, 27, 31, 17, 22,…
$ PROX_CBD <dbl> 7.941259, 6.609797, 6.898000, 4.038861, 11.783…
$ PROX_CHILDCARE <dbl> 0.16597932, 0.28027246, 0.42922669, 0.39473543…
$ PROX_ELDERLYCARE <dbl> 2.5198118, 1.9333338, 0.5021395, 1.9910316, 1.…
$ PROX_URA_GROWTH_AREA <dbl> 6.618741, 7.505109, 6.463887, 4.906512, 6.4106…
$ PROX_HAWKER_MARKET <dbl> 1.76542207, 0.54507614, 0.37789301, 1.68259969…
$ PROX_KINDERGARTEN <dbl> 0.05835552, 0.61592412, 0.14120309, 0.38200076…
$ PROX_MRT <dbl> 0.5607188, 0.6584461, 0.3053433, 0.6910183, 0.…
$ PROX_PARK <dbl> 1.1710446, 0.1992269, 0.2779886, 0.9832843, 0.…
$ PROX_PRIMARY_SCH <dbl> 1.6340256, 0.9747834, 1.4715016, 1.4546324, 0.…
$ PROX_TOP_PRIMARY_SCH <dbl> 3.3273195, 0.9747834, 1.4715016, 2.3006394, 0.…
$ PROX_SHOPPING_MALL <dbl> 2.2102717, 2.9374279, 1.2256850, 0.3525671, 1.…
$ PROX_SUPERMARKET <dbl> 0.9103958, 0.5900617, 0.4135583, 0.4162219, 0.…
$ PROX_BUS_STOP <dbl> 0.10336166, 0.28673408, 0.28504777, 0.29872340…
$ NO_Of_UNITS <dbl> 18, 20, 27, 30, 30, 31, 32, 32, 32, 32, 34, 34…
$ FAMILY_FRIENDLY <dbl> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0…
$ FREEHOLD <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1…
$ LEASEHOLD_99YR <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ LOG_SELLING_PRICE <dbl> 14.91412, 15.17135, 15.01698, 15.26243, 14.151…
$ MLR_RES <dbl> -1489099.55, 415494.57, 194129.69, 1088992.71,…
$ Intercept <dbl> 2050011.67, 1633128.24, 3433608.17, 234358.91,…
$ y <dbl> 3000000, 3880000, 3325000, 4250000, 1400000, 1…
$ yhat <dbl> 2886531.8, 3466801.5, 3616527.2, 5435481.6, 13…
$ residual <dbl> 113468.16, 413198.52, -291527.20, -1185481.63,…
$ CV_Score <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ Stud_residual <dbl> 0.38207013, 1.01433140, -0.83780678, -2.846146…
$ Intercept_SE <dbl> 516105.5, 488083.5, 963711.4, 444185.5, 211962…
$ AREA_SQM_SE <dbl> 823.2860, 825.2380, 988.2240, 617.4007, 1376.2…
$ AGE_SE <dbl> 5889.782, 6226.916, 6510.236, 6010.511, 8180.3…
$ PROX_CBD_SE <dbl> 37411.22, 23615.06, 56103.77, 469337.41, 41064…
$ PROX_CHILDCARE_SE <dbl> 319111.1, 299705.3, 349128.5, 304965.2, 698720…
$ PROX_ELDERLYCARE_SE <dbl> 120633.34, 84546.69, 129687.07, 127150.69, 327…
$ PROX_URA_GROWTH_AREA_SE <dbl> 56207.39, 76956.50, 95774.60, 470762.12, 47433…
$ PROX_MRT_SE <dbl> 185181.3, 281133.9, 275483.7, 279877.1, 363830…
$ PROX_PARK_SE <dbl> 205499.6, 229358.7, 314124.3, 227249.4, 364580…
$ PROX_PRIMARY_SCH_SE <dbl> 152400.7, 165150.7, 196662.6, 240878.9, 249087…
$ PROX_SHOPPING_MALL_SE <dbl> 109268.8, 98906.8, 119913.3, 177104.1, 301032.…
$ PROX_BUS_STOP_SE <dbl> 600668.6, 410222.1, 464156.7, 562810.8, 740922…
$ NO_Of_UNITS_SE <dbl> 218.1258, 208.9410, 210.9828, 361.7767, 299.50…
$ FAMILY_FRIENDLY_SE <dbl> 131474.73, 114989.07, 146607.22, 108726.62, 16…
$ FREEHOLD_SE <dbl> 115954.0, 130110.0, 141031.5, 138239.1, 210641…
$ Intercept_TV <dbl> 3.9720784, 3.3460017, 3.5629010, 0.5276150, 1.…
$ AREA_SQM_TV <dbl> 11.614302, 20.087361, 13.247868, 33.577223, 4.…
$ AGE_TV <dbl> -1.6154474, -9.3441881, -4.1023685, -15.524301…
$ PROX_CBD_TV <dbl> -3.22582173, -6.32792021, -4.62353528, 5.17080…
$ PROX_CHILDCARE_TV <dbl> 1.000488185, 1.471786337, -0.344047555, 1.5766…
$ PROX_ELDERLYCARE_TV <dbl> -3.26126929, 3.84626245, 4.13191383, 2.4756745…
$ PROX_URA_GROWTH_AREA_TV <dbl> -2.846248368, -1.848971738, -2.648105057, -5.6…
$ PROX_MRT_TV <dbl> -1.61864578, -8.92998600, -3.40075727, -7.2870…
$ PROX_PARK_TV <dbl> -0.83749312, 2.28192684, 0.66565951, -3.340617…
$ PROX_PRIMARY_SCH_TV <dbl> 1.59230221, 6.70194543, 2.90580089, 12.9836104…
$ PROX_SHOPPING_MALL_TV <dbl> 2.753588422, -0.886626400, -1.056869486, -0.16…
$ PROX_BUS_STOP_TV <dbl> 2.0154464, 4.4941192, 3.0419145, 12.8383775, 0…
$ NO_Of_UNITS_TV <dbl> 0.480589953, -1.380026395, -0.045279967, -0.44…
$ FAMILY_FRIENDLY_TV <dbl> -0.06902748, 2.69655779, 0.04058290, 14.312764…
$ FREEHOLD_TV <dbl> 2.6213469, 3.0452799, 1.1970499, 8.7711485, 1.…
$ Local_R2 <dbl> 0.8846744, 0.8899773, 0.8947007, 0.9073605, 0.…
$ geometry <POINT [m]> POINT (22085.12 29951.54), POINT (25656.…
$ geometry.1 <POINT [m]> POINT (22085.12 29951.54), POINT (25656.…
summary(gwr_sf_adaptive$SDF$yhat)Length Class Mode
0 NULL NULL
7.5.5 Visualizing local R2
The code chunks below is used to create an interactive point symbol map.
tmap_mode("view")
tm_shape(mpsz_svy21)+
tm_polygons(alpha = 0.1) +
tm_shape(gwr_sf_adaptive) +
tm_dots(col = "Local_R2",
border.col = "gray60",
border.lwd = 1) +
tm_view(set.zoom.limits = c(11,14))tmap_mode('plot')